The standard library bind(f,...) requires a sufficiently resolved callable type, the overloads of notDifferent with default parameters prevents selection of the intended overload.

I tried submitting PR brewsci/homebrew-bio/pull/535 for upgrading to 1.9.6, but CI fails on macOS. I guess changes in path string (like summstats vs SummStats) is preventing proper tracking of diffs on case-insensitive file systems. On macOS, unstaged changes show up right after git clone.

git status:

On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git checkout -- <file>..." to discard changes in working directory)

        modified:   Sequence/SummStats.hpp
        modified:   Sequence/SummStats/Garud.hpp
        modified:   Sequence/SummStats/Makefile.am
        modified:   Sequence/SummStats/Makefile.in
        modified:   Sequence/SummStats/lHaf.hpp
        modified:   Sequence/SummStats/nSL.hpp
        modified:   src/SummStats/Garud.cc
        modified:   src/SummStats/lHaf.cc
        modified:   src/SummStats/nSL.cc

git diff Sequence/SummStats.hpp:

diff --git a/Sequence/SummStats.hpp b/Sequence/SummStats.hpp
index a0bda1a1..e9810ac8 100644
--- a/Sequence/SummStats.hpp
+++ b/Sequence/SummStats.hpp
@@ -1,13 +1,23 @@
-#ifndef __SEQUENCE_SUMMSTATS_HPP__
-#define __SEQUENCE_SUMMSTATS_HPP__
+/// @file Sequence/summstats.hpp
+/// \brief Include all summary statistic functions and types
+#ifndef SEQUENCE_SUMMSTATS_HPP__
+#define SEQUENCE_SUMMSTATS_HPP__

-/*! \file SummStats.hpp
-  Header file for summary statistic of variation data.
+/*!
+ *  \defgroup popgenanalysis Analysis of molecular population genetic data
+ *  \brief Summary statistics and other analysis of Sequence::VariantMatrix
+ *  \ingroup popgen
+ *
+ *  See @ref md_md_tutorial.
+ *
 */

-#include <Sequence/SummStats/nSL.hpp>
-#include <Sequence/SummStats/Garud.hpp>
-#include <Sequence/SummStats/lHaf.hpp>
-#include <Sequence/SummStats/Snn.hpp>
+#include "summstats/generic.hpp"
+#include "summstats/classics.hpp"
+#include "summstats/nsl.hpp"
+#include "summstats/nslx.hpp"
+#include "summstats/ld.hpp"
+#include "summstats/lhaf.hpp"
+#include "summstats/garud.hpp"

Some compilers/stdlibs (Intel 15 / libstdc++ 4.6) have different isfinite functions in the global namespace and in namespace std.

Using using statements to put std::isfinite in the global namespace or an global unnamed namespace makes unqualified isfinite(x) calls ambiguous.

Dealing with cmath vs. math.h and where they put their definitions is a proper mess, putting the alias in the closest namespace yields the right result on both GCC 4.9.1 and Intel 15.

Also, stdlib.h needed for malloc/free.

Hi Kevin,

Blimey, great work, way beyond 'it just works'! I have had some minor trouble installing this, so it could already be a small contribution.

Keep up the good work, Richel Bilderbeek

Hi,

I need to compile libsequence because it's a dependency for ms-stat.

I'm on CentOS6 and I use devtoolset-1.1 to have access to gcc 4.7.2 But I got this message during the make step:

make[2]: Entering directory `/usr/local/install/src/genome2/libsequence-1.8.9/src'
depbase=`echo Seq/Fasta.lo | sed 's|[^/]*$|.deps/&|;s|\.lo$||'`;\
    /bin/sh ../libtool  --tag=CXX   --mode=compile g++ -DHAVE_CONFIG_H -I. -I..    -Wall -W -Woverloaded-virtual -Wnon-virtual-dtor -Wcast-qual -Wconversion -Wsign-conversion -Wsign-promo -Wsynth -ffor-scope   -DNDEBUG -g -O2 -std=c++11 -MT Seq/Fasta.lo -MD -MP -MF $depbase.Tpo -c -o Seq/Fasta.lo Seq/Fasta.cc &&\
    mv -f $depbase.Tpo $depbase.Plo
libtool: compile:  g++ -DHAVE_CONFIG_H -I. -I.. -Wall -W -Woverloaded-virtual -Wnon-virtual-dtor -Wcast-qual -Wconversion -Wsign-conversion -Wsign-promo -Wsynth -ffor-scope -DNDEBUG -g -O2 -std=c++11 -MT Seq/Fasta.lo -MD -MP -MF Seq/.deps/Fasta.Tpo -c Seq/Fasta.cc  -fPIC -DPIC -o Seq/.libs/Fasta.o
In file included from Seq/Fasta.cc:24:0:
../Sequence/Fasta.hpp:53:18: error: ‘Sequence::Seq::Seq’ names constructor
In file included from ../Sequence/Fasta.hpp:45:0,
                 from Seq/Fasta.cc:24:
../Sequence/Seq.hpp:58:7: warning: unused parameter ‘seq’ [-Wunused-parameter]
Seq/Fasta.cc: In constructor ‘Sequence::Fasta::Fasta(const Sequence::Seq&)’:
Seq/Fasta.cc:33:43: note: synthesized method ‘Sequence::Seq::Seq(const Sequence::Seq&)’ first required here
In file included from ../Sequence/Fasta.hpp:45:0,
                 from Seq/Fasta.cc:24:
../Sequence/Seq.hpp: At global scope:
../Sequence/Seq.hpp:59:7: warning: unused parameter ‘seq’ [-Wunused-parameter]
Seq/Fasta.cc: In constructor ‘Sequence::Fasta::Fasta(Sequence::Seq&&)’:
Seq/Fasta.cc:37:50: note: synthesized method ‘Sequence::Seq::Seq(Sequence::Seq&&)’ first required here
make[2]: *** [Seq/Fasta.lo] Error 1
make[2]: Leaving directory `/usr/local/install/src/genome2/libsequence-1.8.9/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/local/install/src/genome2/libsequence-1.8.9'
make: *** [all] Error 2

Do you have an idea?

Thanks

I would like to point out that identifiers like “__POLY_TABLE_TCC__” and “__SEQ_SAMFUNCTIONS_HPP__” do eventually not fit to the expected naming convention of the C++ language standard. Would you like to adjust your selection for unique names?

Several functions consuming bamrecords were not guarded by the feature detection macro HAVE_HTSLIB that Sequence::bamrecord's definition was guarded with.

I'm not willing to track htslib development, and prefer to refocus libsequence on pop gen.

• [ ] add a --no-htslib option to configure
• [ ] mark anything using htslib as deprecated.

Hello, Do you know about Conan? Conan is modern dependency manager for C++. And will be great if your library will be available via package manager for other developers.

Here you can find example, how you can create package for the library.

If you have any questions, just ask :-)

FASTQ file are incorrectly read in if there is whitespace in the sequence name, which apparently NCBI/SRA adds, apparently because they don't feel good unless they've modified what people upload...

New unit tests have revealed a bug in this function.

The previous unit tests were too simple.

See #43 for additional background.

This function is the basis for many haplotype statistics, which will all be a bit off as a result.

• [x] nSL/iHS
• [x] H1, H12 et al.
• [ ] Hudson's Snn
• [ ] Fst
• [x] lHaf
• [ ] Wall's B, Q, etc.
• [ ] r^2/D/D'

This will leave us with pretty much everything except the Fu/Li statistics? There are also some from Rozas that msstats did but never made it into libsequence.